Background: Previous studies have demonstrated an association between body mass index (BMI) and survival after breast cancer diagnosis, but the direction and strength of this relationship is inconsistent. Comorbidities that are more common in patients who are obese (BMI≥30) – including diabetes (DM), hypertension (HTN), and hyperlipidemia (HLD) – are often conflated with obesity with regard to their effects on breast cancer outcomes. We sought to determine the effect of BMI on overall survival (OS) among women with breast cancer after controlling for obesity-associated conditions. Methods: Women≥18y diagnosed with stage 0-IV breast cancer at an academic institution from Jan 2014-Jul 2016 and with known BMI at diagnosis were identified. χ² and ANOVA tests were used to compare intergroup differences. BMI was categorized as normal (<25), overweight (25-29.9), class 1 obesity (30-34.9), and class 2/3 obesity (≥35). Unadjusted OS by BMI class was estimated with the Kaplan-Meier method. Cox proportional hazards models were used to estimate the association of BMI with OS after adjusting for covariates including obesity-associated conditions. Results: 1027 patients were included (median follow-up 46.8 mos): 296 (28.9%) were overweight, 227 (22.1%) had class 1 obesity, and 207 (20.2%) had class 2/3 obesity. Non-Hispanic (NH) black women were overrepresented among obese patients, making up 25% (n=257) of all patients but 37.5% of obese patients. Rates of DM, HTN, and HLD increased with increasing BMI (all p<0.01). Unadjusted OS differed significantly by BMI class, with overweight women having the worst 5-year OS (log-rank p=0.02). After adjustment, BMI continued to be associated with OS, with overweight women having significantly worse OS vs normal-weight women, but there was no significant association between obesity and OS (Table). Conclusions: Despite higher rates of DM, HTN, and HLD with increasing BMI, a diagnosis of obesity was not associated with worse OS in women with breast cancer but being overweight was, suggesting the need for a more nuanced understanding of body composition, obesity-associated conditions, and their respective potential impact on breast cancer outcomes.

HR, hazard ratio. *Also adjusted for age, race/ethnicity, biomarker status, and clinical stage; HLD excluded to avoid overfitting.